Why I built
Martia

Excel and me

For years, tracking my finances was simple. I was single, had one bank account, kept a spreadsheet. I knew exactly how much I spent on food, how much on transport, how much was left over. It worked.

Freelancing and currencies

For many years I worked as a freelancer for international clients — invoices in euros, payments in dollars, foreign currency accounts at a Polish bank. The spreadsheet started breaking: exchange rates, conversion dates, currency differences, multiple income streams at once.

To understand how much I actually earned and how much I spent, I had to sit down and manually tally everything up. I knew I should do it regularly — but there was never time. Or I simply didn't have the energy to deal with it.

Then I got married

Joint account. Two incomes. Dogs. A house. My wife pregnant. Suddenly I had no idea where the money was going. As the one in the family who “handles the finances,” I felt like a CFO without any tools.

I tried AI

As a software engineer with 9+ years of experience at Maze, Daftcode, and Dreamdata, I thought: let me just throw my bank statements into Claude. Let AI figure it out.

But the models made mistakes — they misclassified internal transfers between my own accounts as expenses, mixed up currencies, drew wrong conclusions about spending patterns. Every session started from scratch. No memory, no context.

The key insight

AI models are only as smart as the data they receive. You need a transformation engine — one that cleans up bank statements, makes sure the numbers add up, handles multiple currencies and internal transfers. An LLM can't do that in a single chat session. It takes proper data engineering.

So I built it

If I have this problem, every engineer probably does too. Actually — everyone does. I couldn't find a single European tool that does agentic AI financial planning. So I built one.

Even if it ends up being used only by me and my friends — it's worth it. But I believe every one of us deserves a great AI financial planner.